MCPcopy Create free account
hub / github.com/Open-GD/OpenGD / compute_error

Function compute_error

Source/external/fast_float.h:1708–1714  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

1706// the power2 in the exponent will be adjusted by invalid_am_bias.
1707template <typename binary>
1708fastfloat_really_inline FASTFLOAT_CONSTEXPR20
1709adjusted_mantissa compute_error(int64_t q, uint64_t w) noexcept {
1710 int lz = leading_zeroes(w);
1711 w <<= lz;
1712 value128 product = compute_product_approximation<binary::mantissa_explicit_bits() + 3>(q, w);
1713 return compute_error_scaled<binary>(q, product.high, lz);
1714}
1715
1716// w * 10 ** q
1717// The returned value should be a valid ieee64 number that simply need to be packed.

Callers

nothing calls this directly

Calls 1

leading_zeroesFunction · 0.85

Tested by

no test coverage detected